无线传感网络中基于自动学习机的数据传输策略  

Data Transmission Strategy Based on Automatic Learning Machine in Wireless Sensor Network

作　　者：周丰 朱丽军[2] 王璇[1] 宋海金 ZHOU Feng;ZHU Lijun;WANG Xuan;SONG Haijin(Nanhu College,Jiaxing University,Jiaxing Zhejiang 314001,China;College of Mechanical and Electrical Engineering,Jiaxing University,Jiaxing Zhejiang 314001,China;Computer College,Civil Aviation Flight University of China,Guanghan Sichuan 618307,China)

机构地区：[1]嘉兴学院南湖学院,浙江嘉兴314001 [2]嘉兴学院机电工程学院,浙江嘉兴314001 [3]中国民航飞行学院计算机学院,四川广汉618307

出　　处：《传感技术学报》2020年第1期130-133,152,共5页Chinese Journal of Sensors and Actuators

基　　金：国家自然科学基金项目(51805209);浙江省自然科学基金项目(LQ17E050010)。

摘　　要：尽管多径路由能够提高数据包传输成功率,但是其增加参与数据传输的节点数,提升了网络能耗。为此,提出基于自动学习机的单径路由(Single Path-based Learning Automation Routing,SPLAR)。设计SPLAR路由的主旨在于以最少能耗实现数据包传递率的最大化和端到端传输时延的最小化。通过自动学习机构建每个节点的动作集,并依据环境学习更新每个动作的概率。再选择具有最高概率的动作传输数据。仿真结果表明,提出的SPLAR路由在数据包传递率和端到端传输时延具有较好的性能。Although multipath routing can improve the success rate of packet transmission,it increases the number of nodes participating in data transmission and improves the network energy consumption.Therefore,a Single path-based Learning Automation Routing(SPLAR)is proposed.The purpose of SPLAR routing is to maximize packet transfer rate and minimize end-to-end transmission delay with minimum energy consumption.The action set of each node is constructed by the automatic learning machine,and the probability of each action is updated according to the learning environment.Then select the action with the highest probability to transmit data.The simulation results show that the proposed SPLAR route has good performance in packet transfer rate and end-to-end transmission delay.

关 键 词：无线传感网 单径路由 自动学习机 动作集 服务质量 

分 类 号：TPT393[自动化与计算机技术]